AI-Enhanced Data-Driven Approach to Model the Mechanical Behavior of Sustainable Geopolymer Concrete

The increasing environmental concerns associated with Ordinary Portland Cement (OPC) production have driven research towards alternative, sustainable construction materials. Geopolymer concrete (GPC) has emerged as a promising eco-friendly substitute, offering reduced carbon emissions and improved mechanical properties. However, accurately predicting the compressive strength of GPC remains a complex task due to the numerous variables influencing its performance, such as material properties, mix proportions, and curing conditions. This study develops an interpretable machine learning (ML) model to predict the compressive strength of geopolymer concrete, leveraging various ML techniques, including linear regression, decision trees (DT), gradient boosting, support vector regression (SVR), artificial neural networks (ANN), and random forests (RF). To enhance prediction accuracy, a super learner model is employed, integrating these individual techniques. The model's performance is evaluated using metrics such as the coefficient of determination (R²), mean absolute percentage error (MAPE), mean square error (MSE), and root mean square error (RMSE). Additionally, SHAP values and sensitivity analysis are conducted to quantify the impact of each input parameter on the predictions, ensuring the model's transparency and reliability. The proposed approach provides a robust framework for accurately forecasting the compressive strength of geopolymer concrete, thereby contributing to the advancement of sustainable construction practices.